Influence of vegetation type and topographic position on volumetric soil water content dynamics and similarity among surface and deep soil layers

Clarifying the mechanisms governing volumetric soil water content (VSWC) dynamics in soil profiles is essential, as it can help to elucidate soil water transport processes and improve the prediction accuracy of soil hydrological processes. Using Spearman's rank correlation and wavelet coherence analysis methods, similarity in soil profile VSWC dynamics and factors governing VSWC soil profile dynamics in upslopes and downslopes under three vegetation types (evergreen forest [EG], secondary deciduous forest mixed with shrubs [SDFS], and deforested pasture [DP]) at different time scales (hourly, daily, weekly, and monthly) and in different seasons were analyzed. The results revealed significant similarity in the VSWC of different soil depths (P < 0.01), with the similarity decreasing in accordance with the increment in soil depth. Greater VSWC similarity was found in EG than SDFS and DP sites and in upslope than downslope areas at both forest sites. The average significant coherence area (SCA) of VSWC similarity among surface and deep soil layers varied with the time scale, which was in the order of monthly (58.6%) > weekly (42.8%) > daily (21.8%). The effects of soil properties (e.g., texture, saturated hydraulic conductivity), rainfall, and potential evapotranspiration (ETp) on VSWC similarity were related to the time scale and season in which VSWC monitoring took place. Soil properties had apparent effects on VSWC similarity at longer time scales (i.e., monthly), with a high SCA. In contrast, the effects of rainfall and ETp on VSWC similarity were concentrated at weekly and daily scales, with a relatively low SCA. Rainfall and ETp dominated VSWC dynamics in the summer and fall, respectively. These results imply the use of measured VSWC at one soil depth to predict the VSWC at other soil depths was a reliable method. While the influence of time scale effects and seasonal variations on prediction accuracy of VSWC should be considered.